Visual Supernovae Searching at Siding Spring.

(Robert Evans)


Preliminary results are here presented arising from the use of the 40 inch telescope at Siding Spring Observatory for visual supernova hunting over a period of about 18 months. The use of the telescope in this way is continuing. These results are compared with the performance of my 41 cm. Backyard telescope over a similar period of time, and with recently announced results from the Perth Observatory's Automated Supernova Search using their 61 cm. Telescope over a three-year period.


For some years until the beginning of 1986 a visual search for bright supernovae was conducted by the writer with a second-grade quality 25 cm. Telescope from various locations. Twelve supernovae were eventually found with this telescope, and several others were observed which had been found first by others. Since that time, the search has benn continued using a 41 cm. Telescope. Sixteen supernovae have been discovered with the 41 cm. Telescope, and about the same number of other supernovae have been observed with it. The proportion between those discovered and those merely obserced has changed because the competition for discoveries has become greater in recent years. Since 1986, part of the competition has been in the form of professional computerised automated searches using CCDs which have been conducted with larger telescopes, in particular, a 76 cm. Telescope at Berkeley, and now a 61 cm. Telescope at Perth. Several small-scale operations at other locations have also come into existence recently. At the Christchurch meeting of the Astronomical Society of Australia in 1993 I presented a paper which tried to compare results of the Berkeley automated search from 1986 to mid-1991 (Muller et al. 1992) with my 41 cm. Search over the same period. A total of twenty supernovae were seen at Berkeley compared to seventeen seen from my home. (Evans 1994). A difference of only three supernovae in the two searches over a period of nearly six years raised the possibility taht, if a visual search had been conducted with a telescope like the one used at Berkeley, and assuming that the same amount of telescope time was available which the Berkeley team enjoyed, perhaps similar results might have been obtained by the visual search, or it might have been even more fruitful. In order to test these ideas, a funding proposal through the Australian Research Council was launched for a dedicated one-metre telescope in 1994, but this was not successful. This was followed by applications for observing time on the 40 inch telescope at Siding Spring. The Director of Mount Stromlo Observatory, Prof. Jeremy Mould, and the Time Allocation Committee, have supported this plan to the best of their ability, in the face of steady professional competition for use of the telescope. This support has been particularly noteworthy because the search was visual, and did not use the regular CCD equipment. It was also totally staffed by amateur astronomers.


The aim has been to produce statistics of visual supernova searching with one-metre class telescope which could be compared with what can be achieved using a smaller telescope, and also with the performance of a professional automatic search. The method is simply to observe as many different galaxies as possible as regularly as possible, watching for supernovae. At least two people have been needed all the time on the forty inch telescope, one to operate the telescope and one to observe. Those who gave more of their time than others were John Shobbrook, Tom Cragg and Samantha Beaman. Optically, we found that this telescope is an excellent instrument for this work. The most difficult and time.consuming part of the operation was locating the next object. This arises the particular control system which is presently used with this telescope, related to the fact that it is thirty years old. We took oue share of bright moonlight and bad weather, which naturally reduce the potential of the large telescope visual search. This highlights the desiderability of using a dedicated telescope, which would allow darker nights and better conditions to be chosen.


The search commenced with four nights at the end of 1994. During 1995 a total of 52 nights were allocated. So far, during 1996 (up to July 24), a total of 28 nights have been allocated to this project. This included a record night when 379 galaxies were observed in just over eight hours. I was absent on long-service leave during the period between mid-May and mid-July 1996, living at a light polluted site with only a 31 cm, telescope, and therefore missed the supernova in NGC 1084 which was found in Japan in July two months after maximum light. But, I recovered slightly from this with the discovery of SN 1996al in NGC 7689 with team members using the forty inch telescope on July 22, soon after returning home. The number of different galaxies included in all this observing is not known, but probably around 2000.


Within this same period of time (Dec. 18, 1994 to the end of April 1996.), the 41 cm. Backyard telescope played a purely supportive and supplementary role. There is no need to observe any galaxy with this telescope if it has recently been observed with the 40 inch telescope, or was observed as part of the U.K. Schmidt Supernova Search. With the 41 cm. Telescope, all observing in done under fully dark conditions. The working magnitude limit is generally about 15.7. The observing conditions near the centre of the township of Coonabarabran are not as good as those enjoyed on Siding Spring Mountain. Despite this secondary role, a total of 13.267 observations were made in a period of 231 hours of observing. This made an average rate of galaxy observation of 57.4 galaxies per hour throughout the entire period. Of the 41 cm. discoveries, SN 1996X in NGC 5061 was by far the brightest found with either telescope, or by the Perth Observatory group. Of the 40 inch discoveries, SN 1995G in NGC 1643 was also just visible in the 41 cm. telescope, using high magnification, when I already knew where to look for it. But it would never have been discovered with this telescope, as the galaxy was considered to be too distant, and it was not monitored. So, this supernova has not been included in the results of the 41 cm. search.


This search has been conducted over the last three years by Andrew Williams and Ralph Martin. The 61 cm. telescope is now fully automated, although an observer has to be present to start up, and to watch the weather. The magnitude limit during the search was 16.5 - 17.0. (Williams 1996.) The figures cover a three-year period starting early in 1993. Figures for either eighteen-month period are not known, although more work was done in the second half than in the first. Each observation normally took four minutes, thus averaging fifteen galaxy observations per hour. A total of 5530 observations were made. The galaxies searched were 300 expecially chosen large spiral in which a higher supernova rate was expected. Originally, there was an expectation that the search would use up to 50% of available observing time, but, in practice, many other programmes were also catered for. These observations yielded FIVE confirmed supernova discoveries, although the Perth observeres believed they found a sixth supernova which could not be confirmed independently before the galaxy disappeared behind the sun. No other supernovae were seen. It should be noted that the number of supernovae seen is the true guage of the effectiveness of a search, in relation to the number of galaxies searched, and not the number of discoveries. The number of these supernovae which are discovered by any group depends upon what competition there is for discoveries at the time, and this way vary from week to week. It has been objected that the number of supernovae seen is a true guage of effectiveness of an automatic search, but not of a visual search, because a visual observer would necessarily be biased in favour of observing galaxies in which it was known that a supernova had been found, and this would inflate the number of supernovae seen in a visual search. This objections assumes that the bias does not also occur in an automatic search. In theory, the target galaxies for an automatic search are all predetermined. In practice, however, the vagaries of the weather, and the curiosity of the observer, make this assumption into more a dabatable point. Ultimately, the objection raises a question about the way I have chosen the target galaxies for a visual search. The number of galaxies covered each month is considerable. In my mind the galaxies are filed in groups, according to their location in the sky. And they are observed in groupings, also. Any group will normally be observed completely. Ever since 1980, whenever another observer found a supernova in a galaxy which I had been observing on a regular basis, I asked myself "Would I have looked at this galaxy, as part of my normal routine, during the time that the supernova would have been visible in my telescope ?" Almost without exception the answer "Yes.", or "No.", to this question has been very simple. Particularly when using the 41 cm. telescope tha range of galaxies that I observed from month was fairly clearcut. The istances when this range was reduced by my work schedule, or by the weather, did not affect my ability to answer this question about the particular galaxies in which a supernova appeared. The statistics based upon my observations which have been published in the past about the rates at which supernovae appear in the different kinds of galaxies (for example, van den Bergh and McClure 1994) depend upon my ability to answer this question. With the forty-inch search, the range of galaxies covered each observing period was affected by a different range of factors. But only a very small number of galaxies had supernovae in them about which this question needed to be asked, and only for brief periods. So, in practice, it was not a difficult task to answer "Yes.", or "No." About each of the galaxies. But, in any case, the number of "supernovae seen" in a visual search can be listed as an upper limit, if desired.


The Perth search is finding some intrinsically fainter supernovae in the brighter galaxies which were not seen, or were otherwise missed, in the visual search. The visual search can miss supernovae within one magnitude of the limit (16.5 - 17.5) through having to observe in moonlight, a factor which does not degrade CCD images in the same ways. In the visual search, objects at the very faint end have been missed on two occasions that are known, and may have occurred on other occasions. NGC 908 was observed by us a few nights after the Perth group had found SN 1994ae estimated at mag. 16.5. While I knew that a supernova was present I did not know its location within the galaxy, and no new star was seen. Also, NGC 1398 was observed by us several days before the Perth group found SN 1996N at mag. 16.5. In the first instance, I believe that the supernova was fainter visually than the (red) CCD estimate published. As a result it was not seen by us. In the second instance, the star was seen, but was not recognised as a new object. So, our failure to recognise the new object on this occasion should be put down to human error. The Perth search's smaller number of observations per supernova seen (922-1018-2211) may be due to the particular selection of target galaxies used by them. The visual search covered a much wider range of galaxies, both in type and luminosity. This created a sample of galaxies with a much lower expected supernova rate than the Perth sample. Allowance must also be made for the fact that 40 inch and 41 cm. figures are only for eighteen months, whereas the Perth figures are for three years. On the average, twice as many supernovae are available to be discovered or seen in a three-year period than in eighteen months. A surprising result is that the 40 inch search saw a supernova every forty-one hours, and the 41 cm. search every thirty-eight and a half hours, whereas the Perth computer saw one every sixty-one and half hours. This difference probably arises from the great difference in the average speed of observing, which, in turn, allowed a much wider range of galaxies to be seen by the two visual searches. We were quite surprised that the 40 inch was able to observe galaxies at least sixty percent faster than the average speed of the Perth search, despite the fact that the 40 inch does not have the advantage of computerised location of targets. The efficiency of the automatic search could be improved by using faster CCD and computers. Large telescope visual searches can also be improved by the computerised location of target galaxies. In the case of the 40 inch telescope at Siding Spring, the use of the Cassegrain focus port provides a much quicker use of the telescope than would be possible if a standard Newtonian configuration was being used. A much longer list of different galaxies needs to be observed regularly with the 40 inch, compared to the smaller telescope, if the number of supernovae to be seen is to be improved, and the proper potential of the larger telescope is to be realised. The fainter magnitude limit which is possible with the 40 inch telescope should also be exploited more fully. This, again, emphasises the value of, and need for a dedicated telescope for this kind of observing.


The value of visual searching has again been demonstrated, and support for this kind of work should be improved. Especially, visual searching with more modest telescopes is more useful in finding brighter supernovae, which, in turn, have a particular value for scientific research at present. However, at this time in the history of astronomy, when supernova studies are playing an important role in so many areas of astrophysical interest (Cappellaro 1996), any efficient method of finding new supernovae deserves support. This centainly applies to visual searches using larger telescopes such as the forty-inch, or other telescopes of similar capacity, which can monitor a large number of galaxies, and over considerable distances. These figures also indicate that the Perth search was very effective, and probably found all the supernovae that fell within the parameters of its set task. Both types of searching have their own particular merit and role to play. Despite some claims to the contrary, there is no indication that any one method will make the other methods pf searching unnecessary. None of the three searches described here would have found all of the supernovae listed. Use of this dedicated telescope is a great advantage, especially in a visual search, where access to moonless nights and good weather is more important. It is especially desirable that a dedicated and user-friendly telescope of substantial aperture should be available to prosecute this work more fully and efficiently. The results obtained with the 40 inch telescope were made only one part in seven of allocatable observing time, including the good percentage of bright moonlight. So, a great deal more could be achieved if more observing time was available, and darker nights.


I am very grateful, in the first place, to Dr. Russel Cannon, until recently Director of the Anglo-Australian Observatory, for his support of this project throughout its history. I am also grateful to the Director of the Mount Stromlo and Siding Spring Observatory, Prof. Jeremy Mould, and the Time Allocation Committee, for their continuing support, and for the privilege of extensive use of their telescope without financial cost to us. My helpers at the telescope have been numerous. Those who have given more of their time than others are John Shobbrook and Tom Cragg, both of Coonabarabran, and nineteen-year-old Samantha Beaman of Pimpama, Queensland.